r 2006 British HIV Association HIV Medicine (2006), 7, 487–503

British HIV Association (BHIVA) guidelines for the treatment of HIV-infected adults with antiretroviral therapy (2006)

B Gazzard on behalf of the BHIVA Writing Committee*

Introduction attributed to staffing but this also included $72 per annum for patient incentives, which are unlikely to be deployed in This summary document is an update to the full British HIV the UK. Although it is difficult to extrapolate these costs to Association (BHIVA) Treatment Guidelines published in NHS services in the UK, the absolute cost of adherence HIV Medicine in July 2005 (Volume 6, Supplement 2). Only interventions appears to be relatively low, and slightly less the ‘What to start with’ and ‘Treatment-experienced than that of a genotype resistance test [5]. patients’ sections have been completely rewritten. The tables of recommendations (Tables 1–7) have also been updated to include new data. Please refer to the full When to start treatment guidelines for more information. Primary HIV Adherence No new data have emerged in the past year to change the view of the Committee that, if antiretroviral therapy is Since the last version of these guidelines there have been started in the context of primary HIV infection, it should few new data relating to adherence to in normally only be in the setting of a clinical trial (e.g. the general, or to highly active antiretroviral therapy (HAART) MRC SPARTAC study [6]). in particular. Studies continue to support a role for brief interventions including treatment-related education and individualized adherence problem-solving in increasing Established HIV infection adherence (at least over the following months) [1,2]. A Decisions on when to start treatment are based on an study looking at using long-term partners to support assessment of the risk of disease progression over the adherence showed a marked beneficial effect on adherence medium term if treatment is not started (e.g. using data in the short term, but this was not sustained at 6 months from the CASCADE collaboration [7]), vs. the potential [3]. In the Gilead 934 study, although adherence (measured risks of starting treatment earlier (toxicity and resistance). by pill count) was statistically significantly higher in the The risk of disease progression is determined largely by the once-daily arm vs. twice-daily, the clinical difference was CD4 cell count, plasma viral load and the age of the patient. small (3%) and may have arisen because of the poorer No new data have emerged since the previous guidelines to tolerability of the twice-daily arm rather than because of alter the Committee’s view that treatment should start any benefit of once-daily dosing [4]. before the CD4 cell count has fallen to below 200 cells/mL. Research in the USA examined the costs in real terms of This is because of the higher risk of death or disease delivering a variety of adherence interventions over the progression in patients who delay starting treatment until first year of patient support. The majority of interventions their CD4 cell count has fallen to this level, and the reduced were delivered by nurses. Median direct annual cost was efficacy of HAART when introduced at a count of $420 (range $60–700) per patient; 66% of this was o200 cells/mL in some studies. The implication of this is that earlier diagnosis of infection remains paramount; this

Correspondence: Professor Brian Gazzard, Chelsea and Westminster is only likely to be achieved by much more widespread and Hospital, 369 Fulham Road, London, SW10 9NH, UK. normalized use of HIV antibody testing [8]. In patients with e-mail: [email protected] counts above this level, the risk of disease progression *BHIVA Writing Committee members and contributors: EJ Bernard, M Boffito, D Churchill, S Edwards, M Fisher, AM Geretti, M Johnson, C Leen, without therapy should be considered. Therapy should be B Peters, A Pozniak, J Ross, J Walsh, E Wilkins and M Youle started if the patient wishes to start, the medium-term risk

487 488 B Gazzard

of disease progression is high (e.g.42–4% in 6 months) What to start with and the likelihood of harm as a result of drug toxicity or There is overwhelming evidence from cohort studies that the selection of resistant virus is considered to be low. This the very dramatic fall in AIDS-related mortality and is likely to mean that the majority of patients with CD4 frequency of AIDS events seen in the developed world counts 4350 cells/mL should continue to defer treatment. over the last 10 years coincides with the introduction of Although cohort studies [9–11] have consistently shown that HAART. Further improvements in treatment have led to the risk of disease progression in this group tends to be lower fewer naı¨ve patients now failing first-line therapy. foruptoseveralyearsiftherapyis started earlier, this is at the Measurement of the success of a regimen is achieving a cost of significant drug toxicity, necessitating drug switching viral load of o50 HIV-1 RNA copies/mL within 3–6 or stopping therapy in a significant minority [9], and the months of starting therapy and then maintaining this absolute risk of disease progression in this group remains thereafter. Regardless of the baseline viral load, a level of very low. Other cohort studies (albeit in patient groups using 1000 copies/mL has been found to be achievable in the regimens that might not be recommended currently [12,13]) majority of people by 4 weeks from start of therapy. Failure have demonstrated that selection of drug-resistant virus to achieve this is strongly associated with failure to reach occurs in 25% of patients after 2–3 years, suggesting that viral load below 50 copies/mL within 24 weeks. A viral load deferral of initiation may preserve future treatment options. above 1000 copies/mL measured 4 weeks after the initiation Patients randomized to the drug conservation (DC) arm of therapy therefore should prompt questions over possible (i.e. intermittent therapy) of the SMART study [14] had a poor adherence or other reasons such as reduced drug significantly increased risk of opportunistic infection or levels or primary drug resistance. death after 16 months of follow up compared to those in HAART regimens should be individualized in order to the virological suppression (VS) arm (i.e. continuous achieve the best potency, adherence and tolerability; to therapy). Importantly, this benefit was seen across all minimize potential long-term toxicity and to avoid any CD4 cell count strata, not only in those with a nadir CD4 likely drug–drug interactions. The cost of the regimen should count of o350 cells/mL. These data have been interpreted also be considered. as evidence that drug treatment should be started at an earlier stage than is currently the case. However, analyses of the relative rates of disease progression in the two arms Which HAART regimen is best? indicate that the absolute risk of disease progression remains relatively low in the DC arm of the study Previous guidelines have been hampered by the paucity of (approximately 2% per year compared to 1% in the VS published data to definitively inform whether HAART arm). This is consistent with existing data from cohort based around a non-nucleoside reverse transcriptase studies, and as such does not provide convincing evidence inhibitor (NNRTI) or boosted protease inhibitor (PI) regimen to change the recommendations of the 2005 guidelines. is preferable. Moreover, there have been only limited data Rather, this reinforces the evidence that there is a on comparative efficacy among boosted PIs. continuous increase in the risk of disease progression and Studies have shown the superiority of (EFV)- death in untreated patients as the CD4 cell count falls, and containing regimens over nelfinavir (NFV)-, boosted saqui- that decisions on starting therapy need to be individua- navir (SQV)- and boosted -containing regimens lized, taking into account the specific circumstances and [16–18]. A recently presented controlled trial comparing a risks present in each patient. Patients in the DC arm were currently recommended boosted PI [-boosted lopi- also found to have much higher rates of cardiovascular, navir (LPV/r)] with EFV in naı¨ve patients has demonstrated renal and hepatic disease than were seen in the VS arm. If on intention-to-treat (ITT) analysis the superiority of an this phenomenon is attributable to the effect of untreated EFV-based HAART regimen over 96 weeks [19]. This HIV infection, this would be a further argument for earlier resulted from a lower rate of virological failure in the EFV therapy. As the tolerability and toxicity profiles of initial arm, with no significant difference between the arms in therapy improve and the risk of selecting resistant virus treatment-limiting toxicity. Less class-emergent resistance with initial therapy falls [15], it will increasingly become and improved immunological response was, however, seen reasonable to consider starting therapy earlier than in the boosted LPV arm. In addition, a 48-week study previously (i.e. earlier in the CD4 count range 200– comparing LPV/r with ritonavir-boosted 350 cells/mL, and above counts of 350 cells/mL in selected (FOS/r) has demonstrated the noninferiority of twice-daily patients, such as older patients, those with higher viral FOS/r at all CD4 cell count and viral load strata [20]. load, and those with other comorbidities such as hepatitis B It is important to select the regimen best suited to the or C virus infection). individual patient, and therefore to fully assess baseline

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 489

risk factors for resistance, hepatitis B/hepatitis C virus in some countries/groups, increasing incidence of primary coinfection, cardiovascular disease, diabetes, and psychia- resistance. A single mutation is sufficient to confer tric disease. In addition, lifestyle issues including smoking, resistance to EFV (leading to the loss of currently licensed obesity, alcohol use, and recreational drug use should be drugs in this class in future regimens), although not to the taken into account, as well as gender and the potential for second-generation NNRTIs currently under evaluation in pregnancy. The advantages and disadvantages in terms of Phase III/IV studies. NNRTI resistance is also almost always potency, convenience, toxicity, salvageability and potential accompanied by the emergence of nucleoside mutations, drug–drug interactions are discussed below. limiting options for this class as well. In view of these recently presented data, the Committee The major side effect of EFV is dysphoria, which needs to recommends that EFV-based HAART should ordinarily be be discussed in detail with the patient prior to commencing the first-line choice for naı¨ve patients. However, it must be the drug. Manifestations include vivid dreams and/or emphasized that any drug combination must be tailored to nightmares, sleep and mood disturbance, drowsiness and the patient and, in many instances, including primary disorientation. Most are mild–moderate and self-limiting reverse transcriptase resistance, a boosted PI is preferable. and can be managed with a short course of hypnotics: it is This therefore remains a recommended alternative option. unusual for patients to discontinue the drug for this reason With a second wave of drugs within the PI and NNRTI within trials. Nevertheless, in a small minority, symptoms classes as well as novel target classes currently in Phase III/ persist and may be severe enough to warrant switching to IV studies, we believe that patients should continue to be an alternative agent. The evidence is conflicting as to informed about and encouraged to participate in available whether or not side effects are more common in individuals clinical trials to further answer these questions. with a previous psychiatric disturbance [25]. Rashes do occur, but severe rashes with EFV are unusual (the incidence of Stevens–Johnson syndrome is 0.1%). Similarly, hyper- Two nucleoside reverse transcriptase inhibitors (NRTIs) sensitivity hepatitis occurs (3.4% in the 2NN study) but plus an NNRTI fulminant hepatitis is exceptionally rare. Lipid abnormal- NNRTI-based regimens have now been extensively studied ities, mainly rises above baseline values in total and low- with good durability data available for EFV beyond 3 years density lipoprotein (LDL) cholesterol, are not infrequently [21,22]. They have the advantages of convenience and lack of observed in patients on EFV-containing combinations [23]. long-term toxicity and EFV is the preferred NNRTI for initial EFV may be teratogenic and there have been four therapy. However, EFV is not a drug that is well tolerated by retrospective reports of neural tube defects in mothers taking all patients and the side-effect profile needs to be carefully EFV in the first trimester. Such defects have not been described explained to each patient before starting therapy. Three per in prospectively collected data and the relative risk of its use in cent of patients may experience extreme disorientation early pregnancy remains uncertain [26]. Nevertheless, women including paranoia, nightmares and suicidal ideation, and of childbearing age should be warned about becoming discontinuation rates of 10–20% have been reported over pregnant whilst on EFV and, wherever possible, it should be time in clinical practice. (NVP) is the alternative avoided in women who may contemplate pregnancy. NNRTI but is associated with a worse toxicity profile. EFV has a long half-life when compared with nucleo- sides and it is important to maintain viral suppression for EFV (preferred regimen) a period after discontinuation to prevent functional In many randomized controlled trials, EFV has demon- monotherapy and the emergence of resistance. This can strated improved virological efficacy when compared to be achieved by either continuing the nucleosides or other treatment regimens including PI-, boosted PI-, NVP- substitution of a boosted PI. The optimal duration of and 3NRTI-based regimens. staggered discontinuation is undetermined but is longer in In the 2NN study [23,24], data were obtained to compare certain ethnic groups and may range from 1 to 3 weeks. EFV and NVP in a randomized manner and showed that they were comparable in potency. Equivalence was not NVP (alternative regimen) formally proven, with a small chance that NVP was superior As discussed above, NVP has been compared with EFV in to EFV and a greater chance of the reverse. However, the the 2NN study and has been shown to be of comparable principal reason for the recommendation of EFV as the potency. In this study, however, there was more serious preferred NNRTI is related to toxicity in the NVP arm. toxicity in the NVP arm, with two drug-related deaths, one The major limitation of EFV, as for all currently available from liver failure and one from methicillin-resistant NNRTIs, is the low genetic barrier to resistance. This is Staphylococcus aureus septicaemia in a patient with becoming a major concern because of the continuing and, Stevens–Johnson syndrome.

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 490 B Gazzard

The major side effects are rash and hepatitis. The rash is burden and dosing frequency, and optimizing adherence. It usually mild and self-limiting but may occasionally also limits the development of resistance. The disadvantage manifest as Stevens–Johnson syndrome (incidence 0.3%) of this approach is a possible risk of greater lipid with rare fatalities. The rash is not reduced by the abnormalities, particularly raised fasting triglycerides. coadministration of steroids, which should be avoided In randomized studies, the development of resistance in [27]. Hepatitis is an infrequent side effect that occurs in the patients failing therapy has been shown to be higher in those first 6 weeks of therapy but fulminant liver failure and starting with NNRTI or unboosted PI regimens compared deaths have been reported. Recent analyses have shown a with those starting with boosted PIs [28,29]. This was evident 12-fold higher incidence of symptomatic hepatic events in in the class comparison study between boosted LPV and EFV women with CD4 counts 4250 cells/mL (11% vs. 0.9%) and where, of those viruses sequenced in patients with virological a 6-fold higher incidence in men with CD4 counts failure, NNRTI and primary PI mutations were 48% and 4%, 4400 cells/mL (8% vs. 1.2%) and this drug should be respectively, with NRTI mutations being twice as frequent in avoided in these patients, as well as those with active the NNRTI arm (33 vs. 15%) [19]. The Committee recom- hepatitis B or C virus infection or patients with elevated mends LPV/r or FOS/r dosed twice daily as the preferred liver function tests at baseline. NVP is currently used twice regimens, with alternatives to include boosted SQV (hard gel a day, but the pharmacokinetics and now clinical trial data capsule or film-coated capsule) or boosted (ATZ). indicate that once-daily dosing is possible, although this leads to more abnormalities of liver function [23]. LPV/r (preferred regimen) Based on these data, NVP is not now recommended as a Data from the original licensing study [30] showed superior preferred regimen in patients starting HAART, but should surrogate marker endpoint for patients using LPV/r when be used in patients in whom other regimens would have compared to NFV, with lower numbers of patients disadvantages (e.g. women desiring to become pregnant discontinuing for side effects. Additionally, patients and possibly those with a previous psychiatric history) and randomized to LPV/r who developed virological failure only within the CD4 cell count recommendations. It had no evidence of PI resistance. The main side effects of remains a well-tolerated drug with no adverse effect on this regimen are lipid abnormalities and gastrointestinal lipids, and, when used within the restrictions above, the side effects, with diarrhoea being the predominant risk of hepatotoxicity is probably extremely low. There are symptom. The new formulation of LPV/r tablet is now also cost benefits of NVP over EFV. A 14-day lead-in period available with lower pill burden, no food requirement and of 200 mg daily should be prescribed before increasing less interpatient pharmacokinetic variability, and does not to twice-daily dosing unless switching directly from EFV. need refrigeration. The same recommendations apply when discontinuing NVP as apply for EFV. FOS/r (preferred regimen) Boosted FOS has been compared as once-daily dosed – not recommended 1400 mg/200 mg to NFV (SOLO) and as twice-daily dosed Delavirdine is currently unlicensed in the UK and is not 700mg/100mg to boosted LPV (KLEAN) in clinical studies in recommended. naı¨ve patients [20,31]. Against NFV, there was no overall difference in the surrogate marker endpoint of o400 copies/ mL (69% in the FOS/r arm vs. 68% in the NFV arm) although Two NRTIs plus a boosted PI significantly more patients with viral loads 100 000 copies/ The dramatic decline in clinical progression and HIV- mL and randomized to RTV-boosted fosamprenavir achieved related deaths followed the introduction of the PI class of viral load suppression. Findings from the KLEAN study show antiretrovirals. These agents have shown clinical and FOS/r to be noninferior to LPV/r (66% of those randomized to surrogate marker efficacy in clinical practice, with a FOS/r and 65% of those randomized to LPV/r had viral loads possible enhanced rise in CD4 cell count over NNRTIs in o50 copies/mL at 48 weeks on ITT analysis). Virological naı¨ve patients. Sustained suppression of plasma HIV-1 failure was low (3.7% in the FOS arm vs. 5.4% in the LPV RNA levels has been observed with more than 7 years of arm) and no major PI mutations were identified. Data from continued immunological recovery in patients treated with the SOLO study have also demonstrated that boosted FOS a boosted PI. There is now enough data to recommend that shows durable responses to 120 weeks in naı¨ve patients with if a PI is chosen as part of an initial HAART regimen then, only one case of emergent PI resistance reported to date. with few exceptions, this should be a boosted agent. This is This regimen has a low pill burden and flexible dosing, because ritonavir (RTV) boosting increases drug exposure with the same tolerability and level of dyslipidaemia as and prolongs the drug half-life, thereby reducing pill boosted LPV. It is licensed in the UK to be dosed twice daily.

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 491

SQV-r (alternative regimen) considered as a starting regimen in very occasional Boosted SQV has demonstrated potency in once- and circumstances, for example informed patient choice based twice-daily schedules. It has been compared as a on likely poor adherence if alternative options are used, or twice-daily dose to RTV-boosted (IDV-r) [32] concomitant medication needed such as for tuberculosis. and LPV/r [33], where approximately one-third of patients Nonthymidine-containing 3NRTI regimens [e.g. ABC/3TC/ were drug-naı¨ve. By the protocol-defined primary endpoint tenofovir (TDF) or (ddI)/3TC/TDF] should not be of time to virological failure, LPV/r was superior to SQV/r used because of unacceptably high rates of virological and SQV/r was superior to IDV/r. With the advent of the failure [38,39]. Currently, no triple-NRTI regimen can be 500 mg film-coated tablet, RTV-boosted SQV has a recommended. However, data suggest that ZDV/3TC/TDF relatively low pill burden (three tablets twice daily), and with or without ABC is a possible option when a PI- or represents a well-tolerated alternative to RTV-boosted LPV NNRTI-based HAART cannot be administered. in naı¨ve patients, with possibly less gastrointestinal toxicity. Again, lipid abnormalities may represent a Choice of 2NRTIs (include nucleoside and nucleotide RTI) potential long-term toxicity issue. The once-daily dose is (Table 4) not licensed in the UK and SQV/r should be used twice daily. This advice may change pending the results of Two NRTIs remain an integral component of HAART with clinical trials. either an NNRTI or an RTV-boosted PI. There is no evidence that a third NRTI provides additional benefits and two-class RTV-boosted or unboosted ATZ (specific groups) NRTI-sparing combinations are associated with more Unboosted ATZ has been shown to have similar efficacy to discontinuations because of toxicity and resistance. There NFV and EFV in three clinical studies [34–36]. A fourth are now seven NRTI analogues ( having been study comparing boosted and unboosted ATZ demonstrated withdrawn), and four NRTI coformulations. The availability no difference on ITT analysis at 48 weeks, although there of two new fixed-dose combinations [Truvada (TDF/FTC) were more virological failures in the unboosted arm [37]. and Kivexa (ABC/3TC)] in addition to ZDV/3TC has led to ATZ/r has also been shown to have similar efficacy to the majority of patients who are naı¨ve to being LPV/r in PI-experienced patients. prescribed one of these as their 2NRTI backbone. With The main advantages of ATZ/r are that the drug is dosed increasing emphasis on short-term tolerability, conveni- once daily and has fewer adverse effects on lipids than ence, absence of long-term toxicity, and cost, there are other boosted PI regimens. Its main side effect is merits and limitations of each coformulated 2NRTI back- hyperbilirubinaemia with or without jaundice, but this is bone, which are discussed below. Pretreatment drug not associated with liver enzyme changes and seldom resistance will also influence the choice of NRTI backbone, results in the need to discontinue treatment. It is not although currently there is no discernible increase in the UK licensed in the UK for naı¨ve patients. A disadvantage is the in the rate of NRTI mutations in recently infected patients. interaction of ATZ with acid-reducing agents, notably protein-pump inhibitors. This is not overcome by RTV Coformulated 2NRTIs boosting, and where alternative agents cannot be used, ZDV/3TC (coformulated as Combivir) is the most studied of atazanavir should be avoided. The Committee feels that use the dual NRTI backbones and has, until recently, been the of this drug should be boosted in naı¨ve patients and most popular 2NRTI combination in the UK [40]. When restricted to those with established cardiovascular risk combined with EFV, it has similar virological efficacy to factors and where a PI is required. This advice may change ABC/3TC at 48 weeks [41]. However, in the Gilead 934 pending the results of clinical trials. study, coformulated ZDV/3TC was compared with indivi- dually dosed (FTC)/TDF with EFV as the third agent. Ninety-six-week results demonstrated a signifi- Three NRTIs cantly higher frequency of patients with virological There is now surrogate marker endpoint data suggesting rebound greater than 400 copies/mL (5 vs. o1%) and that // (ZDV/3TC/ABC, adverse events leading to discontinuation in the ZDV/3TC usually combined as Trizivir) is less potent than the arm (12 vs. 5%) [42]. In this study and the Gilead 903 study combination of two NRTIs with either an NNRTI or a PI. [22], TDF with either 3TC or FTC both dosed with EFV In ACTG 5095, the Trizivir arm was less potent than the showed high and durable virological activity with good other two arms (Trizivir/EFV or Combivir/EFV). This tolerability and minimal long-term toxicity. ABC/3TC is the finding was found both at high and low entry viral loads. third coformulated compound (Kivexa) and has been The Committee now feels that Trizivir should only be shown to give effective lasting virological control. In

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 492 B Gazzard

studies where the third drug was either EFV or a boosted PI, HSR was also reported in 3% of ZDV-treated patients in a the proportion of patients who failed virologically was double-blind study) and the summary of product char- 4–10% at 48 weeks [20,41,43]. For both Truvada and acteristics (SPC) states an HSR rate of 5.4%. Pharmaco- Kivexa, the individual components have been shown to be genetic analysis has identified a close association between bioequivalent with the fixed-dose combinations [44]. Higher HSR and carriage of the class 1 HLAB*5701 allele, which to CD4 cell counts at 48 weeks have been demonstrated with a large extent explains the racially defined differences in both TDF/FTC and ABC/3TC when compared to ZDV/3TC. susceptibility [51]. Absence of this allele has a high negative All three NRTI combinations perform well virologically. predictive value for HSR in predominantly Caucasian Most clinician experience is of twice-daily ZDV/3TC. populations [52]. The prospective use of this test in clinical However, because of the increasing evidence implicating practice has been shown to significantly decrease the rate of longer term use with lipoatrophy, the Committee believes it HSR from approximately 5–8% to o1–2%, although its utility is essential to carefully monitor for this complication in remains to be proven in a randomized clinical trial and in patients who receive ZDV, possibly with dual energy X-ray racially diverse populations. Consequently, many centres are absorptiometry (DEXA) scanning, with a switch to an now routinely screening and avoiding ABC in patients with alternative NRTI as soon as there is any subjective or the HLAB*5701 allele. Although undeniably of significant objective indication of its development. benefit in reducing the likelihood of HSR, the Committee feels TDF with either 3TC or FTC has a better tolerability/ that there are still only limited and nonrandomized data on its toxicity profile than ZDV, including less dyslipidaemia. use in clinical practice and that the emphasis remains with Because of isolated reports of renal tubular damage and an patient counselling and physician vigilance. There is no association between related compounds and nephrotoxi- evidence of longer term toxicity with ABC to date, although city, concern has been raised about possible long-term dyslipidaemia is greater than with ZDV. The L74 V mutation is renal toxicity with TDF. Numerous studies, including seen in o1% at 48 weeks in patients on EFV/ABC/3TC or a clinical trial, observational cohort, and expanded access boosted PI (overall virological failure rate 4.0–9.9%) [20,41]. data, have identified serious renal toxicity in approxi- Both K65R and L74 V mutations can lead to difficulties in mately 0.5% of patients, a rate no different to that observed choice of subsequent NRTI drugs if they develop. with comparator NRTIs [45–47]. However, other studies In summary, these three NRTI backbones offer good have demonstrated a small but significant reduction in antiviral efficacy and can all be recommended. When renal function over time when compared to other nucleo- combined with EFV, TDF/FTC leads to fewer outcome sides [48,49]. It is clear that TDF should be used cautiously failures than ZDV/3TC, a difference driven mainly by ZDV- in patients who have, or are at risk of developing, renal related toxicity in the first 24 weeks. Complicating lipo- disease, including those coprescribed potentially nephro- atrophy with long-term ZDV is also an important factor in toxic agents. Moreover, blood biochemistry and urinalysis choice of backbone NRTIs. Use of ABC/3TC is likely to should be performed prior to initiating TDF with regular increase with prospective screening for HLAB*5701, but a monitoring throughout treatment. However, in this man- negative test does not rule out the possibility of HSR, and the agement context, the Committee feels that TDF is a safe and need for careful counselling and monitoring for ABC HSR effective NRTI. When using TDF as first-line therapy, the remains. Ongoing monitoring for long-term toxicity and K65R mutation has been observed in a minority of patients resistance development with all three fixed-dose combina- receiving EFV/TDF/3TC. In the Gilead 903 study, this tions is important, particularly for TDF/FTC and ABC/3TC, mutation occurred in 2.4% at 48 weeks (overall virological where experience is still limited by comparison to ZDV/3TC. failure rate 9.7%). This was mainly observed in those with The relative costs of the individual NRTIs is becoming an CD4 counts of o50 cells/mL and viral loads 4100 000 increasingly important factor in defining treatment path- copies/mL: K65R has not been observed in patients with ways in naı¨ve patients; these are given in Table 1. pretreatment wild-type virus when receiving LPV/r/TDF/ FTC to 48 weeks and EFV/FTC/TDF to 96 weeks [4,50]. Other 2NRTI combinations The generation of this mutation and M184 V will also be (d4T)/3TC is a well-studied nucleoside combination influenced by the third drug choice (NNRTI or boosted PI). with equal antiviral effectiveness to TDF/3TC and ABC/3TC ABC/3TC is generally well tolerated but a hypersensi- but with significantly greater mitochondrial toxicity, includ- tivity reaction (HSR) may occur in the first 6 weeks (median ing peripheral neuropathy and lipoatrophy [22,53]. Because 11 days from drug commencement) and all patients need of this, d4T is not recommended for initial therapy. TDF/ddI is counselling. It has been identified in approximately 8% of a once-daily two-tablet combination. However, all studies naı¨ve-patient studies and is independent of dosing where this 2NRTI backbone has been used with an NNRTI as frequency. These studies used a case reporting form (where the third agent have demonstrated an unacceptably high

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 493

Table 1 Cost of preferred regimens as per Table 4, showing the monthly Conclusions (30-day) cost as set out in list price (May 2006) 1 VAT at 17.5% in d Most clinicians in the UK favour an NNRTI-based regimen Column A Cost (d) for initial therapy, reserving boosted PIs for later. This is

NNRTI based on the perceived low risk of toxicity, the ease of EFV 245 administration, and the genetic frailty of an NNRTI in NVP 188 patients failing a boosted-PI HAART. The recently pre- PI/r LPV/r 361 sented controlled trial comparing boosted LPV with EFV, ATZ/r1 411 which demonstrated the superiority of an efavirenz-based 2 FOS/r 402 HAART regimen, provides support for this practice. More- SQV/r2 393 w IDV/r2 256 over, there is cost benefit in selecting an NNRTI-based TPV/r4* 734 regimen rather than most boosted PIs, which needs to be PI taken into account. Therefore, the Committee recommends ATZ 300 mg 371 NFV 321 that EFV-based HAART should ordinarily be the first-line NRTI choice for naı¨ve patients. However, for some a boosted PI ABC 261z will be preferable and therefore boosted LPV and FOS T20* 1349 remain alternative recommended drugs. It is anticipated that clinicians will favour one of the three coformulated 1 2 4 r ,r or r indicates the number of ritonavir capsules per day. NRTI drugs. They offer excellent antiviral efficacy with a *For experienced patients only. wNot a preferred regimen but recognized potential cost savings. demonstrable benefit of TDF/FTC over Combivir. They z s Trizivir d635. differ in dosing schedule, tolerability, short- and long-term ABC, abacavir; ATZ, atazanavir; IDV, indinavir; EFV, efavirenz; FOS, fosamprenavir; LPV, ; NFV, nelfinavir; NRTI, nucleoside reverse toxicity, and cost. Above all, choice should be tailored to transcriptase inhibitor; NNRTI, nonnucleoside reverse transcriptase inhi- the individual patient. bitor; NVP, nevirapine; PI, protease inhibitor; r, ritonavir; SQV, ; TPV, tipranovir; T20, enfurvitide.

Column B Column C Treatment-experienced patients

NRTI-1 Cost (d) NRTI-2 Cost (d) Total cost Data from the Health Protection Agency indicate that,

w among treatment-experienced patients, NNRTI resistance ZDV 250 mg 196 3TC /FTC 375/388 has remained relatively stable in recent years, whereas TDF 300 179/192 ABC 261 479/492 NRTI and PI resistance has declined. The number of ddI 400 mg 192 440/453 individuals infected with triple-class-resistant viruses has D4T 40 mg* 201 371/384 s fallen from a peak of around 15% in 1999–2001 to around Combivir 380/393 s Truvada 374 8% in 2003–2004 [57]. s Kivexa 492 439 When to switch *d4T/3TC is as effective as other regimens but more toxic and not a preferred regimen. w150 mg tablets. The goals of treatment for the majority of treatment- ABC, abacavir; ddI, didanosine; d4T, stavudine; FTC, emtricitabine; NRTI, experienced patients have changed, and the new paradigm nucleoside reverse transcriptase inhibitor; 3TC, lamivudine; TDF, tenofovir; ZDV, zidovudine. should be aiming for an undetectable and durable HIV plasma viral load supression, wherever possible, leading to failure rate, with the development of early resistance, which immunological improvement with lack of clinical progres- was more marked in patients with more advanced disease sion and improvement in quality of life. This may not be [54]. There is also potential for TDF to potentiate ddI-related possible for a minority of heavily pretreated patients with toxicity. The combination is not recommended. ddI/3TC or virological failure and multiple mutations. In this setting, ddI/FTC as a 2NRTI combination is well tolerated and the goal should be maintaining quality of life by preserving effective [55]. However, ddI-related restrictions on food and if not improving immunological status. In so doing, we the potential for long-term mitochondrial toxicity make this should still aim to optimize viral suppression and impaired choice less popular. ZDV/ddI was a common 2NRTI viral fitness. If not previously examined, reasons for combination prior to HAART. However, no data exist with virological failure should be explored and managed ZDV/enteric-coated ddI in HAART. Similarly, TDF/ABC and accordingly. ABC/ddI have not been evaluated in naı¨ve patients: none of The key considerations in the choice of HIV therapy in- these 2NRTI combinations can be recommended. clude treatment history, comorbid conditions, tolerability,

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 494 B Gazzard

adherence, drug–drug interactions, resistance testing, the likelihood of constructing a virologically suppressive current and future available treatment options, and the and durable regimen. Several drugs that have activity clinical status of the patient. against currently resistant viruses are in a late phase of In treatment-experienced patients with therapy options development and will be used to create an effective drug there are data suggesting that switching as soon as possible combination. is more likely to be successful in terms of achieving an Even partial virological suppression of HIV RNA undetectable viral load when the CD4 cell count is higher 40.5 log10 copies/mL from baseline correlates with clinical and viral load lower. When there are few drug options, it benefits [61], although this must be balanced with the may be better to maintain the patient on a failing regimen ongoing risk of accumulating additional resistance muta- which maximizes the fitness effect and has minimal side tions. There is good evidence that continuing therapy, even effects. in the presence of viraemia without CD4 cell count increases, reduces the risk of disease progression [62]. Other cohort studies suggest continued immunological and clinical What to switch to benefits if the HIV RNA level is maintained at 10 000– The patient with therapy options 20 000 copies/mL [63,64]. In treatment-experienced patients with therapy options, the Virological failure on continuous antiretroviral therapy physician should construct a new HIV treatment which is associated with variable changes in CD4 cell counts includes at least two (or preferably three) active agents which appear to correlate with the replicative capacity of guided by HIV resistance testing and by the patient’s the virus, its ability to induce apoptosis in both infected previous antiretroviral (ARV) history. The use of an agent and uninfected CD4 cells, and its use of CXCR4 vs. CCR5 from a new drug class is likely to be more effective [56–58]. receptor for entry [65]. There is currently no test that can The available data for (T20) show that it is most predict CD4 cell count responses in individual patients effective when used with other drugs to which the patient continuing a failing regimen. The replication capacity is susceptible, based on resistance testing and antiviral assay provides only a partial measure of the multiple experience. When used as the only effective agent, factors that influence virus fitness and pathogenicity. resistance to it occurs within weeks and a future Once resistant virus has become established as the opportunity for constructing an effective regimen is lost. dominant species, the emergence of further resistance If (TPV) or (TMC114) is used in a mutations detectable by routine genotypic testing is widely similar way, then the regimen is less effective than in believed to occur slowly. In one study of patients with viral combination with other effective drugs such as T20. In the load 4200 copies/mL, the average increase per year in the Toro 1 and 2 combined analysis of LPV/r-naı¨ve patients, number of mutations was 0.5 for reverse transcriptase (RT) the percentage reaching less than 400 copies/mL doubled mutations, 0.2 for major PI mutations and 0.3 for minor PI from 30 to 60% at 24 weeks if T20 was used. In the mutations [66]. Within the SCOPE clinical cohort, however, combined analysis of TPV/r patients in the RESIST study, among persons with viral load 41000 copies/mL while on the percentage reaching less than 400 copies/mL went from stable therapy, as many as 44% accumulated at least one 30 to 54% at 24 weeks if T20 was used, and in the new mutation at year 1, and 30% lost at least one active combined analysis in the POWER1 and POWER2 studies of drug [67]. The highest risk for the emergence of further TMC114/r-naı¨ve patients the percentage reaching less than mutations is in persons who initially have limited 50 copies/mL increased from 46 to 64% at 24 weeks if T20 resistance. Conversely, in some patients with multiple was used [59,60]. mutations a genetic deadlock may be achieved that limits Although these studies had different designs and entry further evolution of resistance. However, the majority of criteria they underline the principle that new regimens patients keep accumulating new resistance mutations, should contain two or more fully active drugs. However, including mutations present only as minority variants that this strategy may not be a realistic option when managing escape detection by routine testing. some highly treatment-experienced patients. A ‘V-shape’ relationship exists between the number of mutations and viral load in the setting of treatment failure. The patient with few or no therapy options: continue, Above a certain threshold, the number of mutations is interrupt or change therapy? associated with increases in viral load, reflecting compen- Continue? In treatment-experienced patients with few or satory changes that improve virus fitness and pathogeni- no therapy options, especially if the CD4 cell count is well city [68]. Variants with compensatory changes may carry maintained, it may be better to wait to change treatment mutations in several regions of the HIV genome and do not until an investigational agent is available that will increase necessarily display further increases in drug resistance [69].

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 495

Table 2 Recommendations for starting treatment

Presentation Surrogate markers Recommendation

Primary HIV infection Treatment only recommended in a clinical trial, or if severe illness (CIV) Established infection CD4 o200 cells/mL, Treat (AI) any viral load CD4 201–350 cells/mL Start treatment, taking into account viral load (BIII), rate of CD4 decline (BIII), patient’s wishes, presence of hepatitis C (CIV) CD4 4350 cells/mL Defer treatment (BIII) Symptomatic disease or AIDS Any CD4 count or viral load Treat (AI)

Please refer to the full guidelines for further information (HIV Med 2005; 6 (Suppl. 2): 1–61).

Table 3 Comparison of boosted protease inhibitors (PIs)

Lopinavir/ritonavir Saquinavir/ritonavir Fosamprenavir/ritonavir Atazanavir/ritonavir

Potency naives 1111 111 1111 111§ Durability data 1111 11 111 11 Convenience 111* 111w 111 1111 Tolerability 11 111 111 1111 Lipid profile 1 1 1 1111 Fat changes profile§ 11 11 11 111 Resistance barrier 1111 1111z 1111 1111z Interaction profile 111 11 111 1 Active against resistant virus 1111 111§ 111 111

1111, excellent; 111, very good; 11, moderately good; 1 , not good. *Tablet formulation. w500 mg tablet. zLimited data in naı¨ve patients. However, the Committee feels that there is sufficient evidence available for boosted PIs to allow careful extrapolation of data. §Limited data available.

In addition, under prolonged drug-selective pressure, side effects. Partially interrupting components of a treat- mutations initially present on separate virus variants can ment regimen may have a role to play in such heavily accumulate on the same viral genome. Such linkage cannot treatment-experienced patients and may reduce toxicity be detected by standard genotype analysis [70]. and potential for drug interactions. Taken together, these observations indicate that con- Data from a small pilot study showed that interruption of tinuing a failing regimen can be deleterious and should be PI treatment was associated with stable HIV RNA levels and reserved only for those patients lacking effective treatment waning of PI mutations. However, viral replicative capacity options and maintained for the shortest period possible and HIV RNA levels eventually started to increase after long- pending the availability of new treatments. term (more than 6 months) observation. In contrast, subjects Treatment interrupt? Until recently, treatment interrup- interrupting their reverse transcriptase inhibitor treatment tions were used as one experimental strategy for these then experienced an immediate rise in HIV RNA. Interest- patients. Four studies confirm that interrupting treatment ingly, most subjects had a subsequent loss of the M184 V in an effort to revert to wild type prior to initiation of a mutation. These results suggest that NRTIs may retain direct salvage regimen is not associated with significant durable antiviral activity against the resistant variant [75]. benefits [71–74]. Instead, it may be associated with a rapid One controlled clinical trial including 3TC in a increase in HIV RNA, loss of CD4 cells or clinical disease subsequent regimen after the development of an M184 V progression. The OPTIMA study, which addresses the issues mutation has shown no benefit, but was discontinued for of the optimal number of drugs and the benefits of treatment futility [76]. However, 3TC retains antiviral activity even in interruption, has now completed accrual. Pending these the face of complete phenotypic resistance [77]. It was results, total treatment interruption cannot be recommended recently demonstrated that withdrawal of 3TC from a in the management of the treatment-experienced patient. failing antiviral regimen led to an average increase of 0.5

Change? Until investigational drugs that are effective log10 in viral load [78]. This adds further support to the against currently resistant virus are available, it might notion that 3TC retains some of its activity even in the be possible to change the regimen and recycle drugs presence of genotypic resistance. If, on the basis of previously used or omit drugs from the treatment that resistance testing, there are more potent NRTIs available, are having little antiviral effect and/or contributing to then these can be used with or instead of 3TC.

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 496 B Gazzard

In those patients with no current active therapy options, Table 4 Preferred regimens (choose one drug from columns A, B and C) 3TC may contribute to a salvage regimen even in the Regimen A B C presence of high-level genotypic resistance. This situation has led to the possibility of treating patients with 3TC Preferred EFV ZDV§ 3TC§z z # monotherapy. This strategy has only been used, however, in LPV/r ABC FTC FOS/r TDF# the setting of treatment failure in patients with the M184 V mutation and CD4 counts of 4500 cells/mL. Fifty-eight Alternative SQV/r ddI treatment-experienced patients harbouring 3TC-resistant Specific groups NVP* wz virus were randomized in a trial comparing the immuno- ATZ ATZ/rwz logical and clinical outcomes of 3TC monotherapy and complete therapy interruption. By week 48 in ITT analysis, *Only when CD4 o250 cells/mL in female patients and o400 cells/mLin male patients. immunological failure (CD4 count falling to o350 cells/mL) wWhere there are established cardiovascular disease risk factors and a or clinical failure (grade B or C clinical event) occurred in protease inhibitor (PI) is required. z Currently unlicensed for naıs¨ve patients in the UK. 69% of persons (20/29) in the structured treatment §Coformulated as Combivir . z s interruption (TI) group and 41% of persons (12/29) in the Coformulated as Kivexa s. #Coformulated as Truvada . 3TC monotherapy group. 3TC monotherapy significantly ABC, abacavir; ATZ, atazanavir; ddI, didanosine; EFV, efavirenz; FOS, delayed CD4 cell count decline and reduced viral load fosamprenavir; FTC, emtricitabine; LPV, lopinavir; NVP, nevirapine; r, rebound compared to TI, and only patients in the TI groups ritonavir; SQV, saquinavir; 3TC, lamivudine; TDF, tenofovir; ZDV, zidovudine. experienced clinical failure. Disappearance of resistant variants was reduced and replication capacity was sig- Testing for transmitted resistance continues to be nificantly lower in the 3TC group, implying a beneficial recommended in all newly diagnosed patients. This effect of impaired viral fitness [79]. includes patients with acute seroconversion, established A judicious selection of maintenance therapy is required infection or infection of unknown duration, regardless of in these patients, guided by resistance testing as well as demographic characteristics. The most appropriate sample considerations of tolerability. Each case should be judged is the one closest to the time of diagnosis. on its merits, but as general guidance: (a) it is preferable to The true risk of superinfection remains to be determined, select drugs to which the patient already shows extensive but may be significant in persons with early infection who resistance; (b) attempts should be made at inducing or engage in high-risk behaviour. Following baseline resis- maintaining resistance patterns known to be associated tance testing, repeat testing is not routinely recommended with reduced viral fitness, including exploiting antagon- prior to starting therapy, although it may be considered in isms between resistance pathways and potential hypersus- selected persons. ceptibility effects; (c) the immunological efficacy of the regimen should be reviewed closely; (d) it should be Resistance in treatment-experienced cohorts understood that standard genotypic (and phenotypic) testing does not necessarily reflect virological changes Cohorts with detectable viral load while on HAART show a that may impact on immunological success and the high prevalence of drug resistance [83]. A trend towards a preservation of future treatment options. declining rate of resistance has been observed in drug- It should be remembered that the strategy of using experienced patients in the UK, as a result of improved incompletely suppressive regimes will always be a short- management of antiretroviral therapy and treatment failure term one. It is only relevant until a regimen likely to [84]. suppress viral replication completely can be found. Resistance testing and interpretation Resistance testing Current assays require a viral load of at least 500– 1000 copies/mL to reliably provide a result. Although it Primary or transmitted resistance is possible to obtain results at levels of only a few There is extensive evidence for the transmission of drug- hundred copies/mL, the genotypes (and phenotypes) resistant variants among populations in Europe, with obtained are not necessarily representative of the dominant especially high rates of resistance to NRTIs and NNRTIs virus species and results can be misleading [85]. [80,81]. Current estimates from unselected cohorts in Routine resistance assays do not detect resistant viruses London indicate a prevalence of 7% overall and 16% present at low levels (o20–30% of the total virus among those born in the UK [82]. population), even if these resistant viruses were previously

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 497

Table 5 Changing therapy on first virological failure (BIII)

Presentation Viral load pattern Recommended action

Inadequate virological response Failure to achieve viral load o50 copies/mL. Consider factors affecting plasma drug levels.* to initial regimen If drug exposure optimal and likelihood of resistance low, consider augmenting treatment regimen. If likelihood of resistance high, consider changing all drugs.

Persistent viral load rebound Viral load 450 and o400 copies/mL. Consider factors affecting plasma drug levels.* w where previously o50 copies/mL Sustained viral load rebound to 4400 copies/mL. Consider changing all drugs if effective option available likely to reduce viral load to undetectable levels. Continue regimen and monitor if no effective option currently available for reasons of drug potency, likely poor adherence or tolerability.z

*Factors affecting plasma drug levels include poor adherence, intolerability, drug interactions and incorrect dosing. wA viral load rebound to 41000 copies/mL will allow resistance testing to be performed. Resistance testing with expert interpretation has been shown to have a beneficial effect on short-term virological response to the subsequent regimen. zThere is a risk of developing further mutations in allowing a patient to remain on a virologically failing regimen, which could limit further options for treatment. Please refer to the full guidelines for further information (HIV Med 2005; 6 (Suppl. 2): 1–61). dominant. Limited data indicate that minority resistant Therapeutic drug monitoring (TDM) quasispecies including NNRTI resistance mutations and Large randomized prospective controlled trials remain a M184 V may affect virological responses [86]. The pre- high priority to refine the population needing therapeutic valence of resistance mutations that impact on virological drug monitoring (TDM) of NNRTIs/PIs and define the responses is not known and may vary for different clinical benefit of this test. TDM has been shown to be mutations and antiretroviral regimens. Although assays beneficial in particular clinical scenarios where drug to detect minority species have been developed, they are concentrations are difficult to predict, such as the manage- complex, not standardized and remain research tools only. ment of drug interactions, the settings of pregnancy and Current resistance assays target the reverse transcriptase, paediatrics, salvage therapy settings when TDM and protease and regions of HIV. There is increasing resistance test results can be integrated, cases of patients evidence that other viral regions, including gag for the PIs with renal or hepatic impairment and transplant patients, and RNAseH for the NRTIs, play a role in drug resistance. In cases of toxicity, and the use of alternative dosing regimens most cases these changes alone are not sufficient to confer whose safety and efficacy have not been established. resistance, but further evidence is awaited. Clinical data supporting the use of inhibitory quotients The interpretation of resistance test results is complex. are limited; however, these appear to be superior in Although informative interpretation systems based on predicting failure compared to drug concentrations or clinical outcome measures have been introduced, none of resistance testing alone in extensively pretreated patients the available interpretation systems is completely accurate commencing salvage regimens. and all are subject to change as more data become available. Interpretation of resistance tests is especially difficult with Metabolic complications new drugs and this problem affects both genotypic and phenotypic resistance assays. Genotypic scores and ‘clinical As the prognosis of HIV infection has markedly improved, cut-offs’ are being determined for a growing number of so has our need to manage long-term morbidity associated drugs, which correlate specific mutation patterns and viral with HIV and HAART. phenotype with virological responses, respectively. Pheno- Abnormalities of lipid homeostasis and fat distribution typic cut-offs may vary depending on the phenotypic assay are likely to assume a central role in guiding choices for used. Most importantly, they reflect the treatment history antiretroviral therapy. This is the result of the growing and resistance patterns of the treatment-experienced awareness of the increase of stigmatization and reduction cohort(s) from which they have been derived and rely of adherence associated with lipodystrophy, especially heavily on complex statistical analysis. lipoatrophy, and of the increased cardiovascular risk There are subtype-specific treatment-associated muta- associated with drug-induced metabolic abnormalities. tions that have unknown effects on drug susceptibility. Few prospective studies address the relative risk of different Overall, however, recognized mutations that confer resis- regimens causing the features of lipodystrophy, although tance in subtype B also cause resistance in non-B subtypes ACTG384 suggests that the risks are greater with PIs than and vice versa [87]. with NNRTIs and greater with d4T than with ZDV [88].

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 498 B Gazzard

Table 6 What to change to after first virological failure: summary of The application of current health promotion approaches to recommendations (BII/IV) our patients should include assessments of cardiovascular risks and their appropriate management according to the most Change all drugs if possible Resistance test recommended recent international guidelines, particularly as it has been suggested that HIV, because of its proinflammatory profile, Initial regimen Options to consider might present a greater risk for the development of 2NRTIs 1 PI 2NRTIs*w 1 NNRTI cardiovascular disease. Nonetheless, other considerations, such (with or without low-dose ritonavir) or 2NRTIs* 1 boosted PIz as toxicity and resistance, may outweigh potential cardiovas- § or 2NRTIs* 1 NNRTI 1 boosted PI cular risks, especially in therapy-experienced patients. 2NRTIs 1 NNRTI Boosted PI 1 2NRTIs*

3NRTIs 2NRTIs*w 1 NNRTI New drugs or boosted PI 1 2NRTIs* or boosted PI 1 2NRTIs* 1 NNRTI§ This section is intended to cover only those drugs that are likely to be licensed before the next iteration of the *Change to two new and active NRTIs guided if possible by resistance guidelines. testing. wThis could lead to rapid development of resistance to NNRTIs if the potential exists for NRTI cross-resistance. (TMC125) zLow-dose ritonavir-boosted PI should be considered if resistance to PIs is not found or limited. This is a second-generation NNRTI which was developed §Studies with low-dose ritonavir-boosted PI 1 an NNRTI have shown good results. because of its activity against viruses resistant to present NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, nucleoside reverse members of the NNRTI class. Monotherapy studies have transcriptase inhibitor; PI, protease inhibitor. Please refer to the full guidelines for further information (HIV Med 2005; 6 demonstrated a viral load drop of greater than 1 log10 over (Suppl. 2): 1–61). a 10-day period in naı¨ve patients with a smaller fall in those with previous NNRTI experience. A Phase II study comparing TMC125 to a PI-based regimen in patients who Table 7 British HIV Association (BHIVA) Guidelines for the use of had failed an initial regimen of two nucleoside analogues therapeutic drug monitoring (TDM) and an NNRTI and were PI-naive was discontinued at week Indication Recommendation 12 because of a higher virological failure rate in the TMC125 arm. Bioavailability has since been improved with Routine use Not recommended. Insufficient data Drug interaction Recommended (BIII) a new formulation. In triple-class experienced patients, Liver impairment Recommended (BIII) TMC125 was shown to produce an approximate 1 log10 Pregnancy Recommended (BIII) drop in viral load at 48 weeks, with 30% of patients having Minimizing toxicity Recommended (BIII) for IDV, EFV and ATZ a viral load less than 400 copies/mL compared with 8% Consider (CIII) for other drugs treated with optimized backbone therapy alone. Antiviral Monitoring adherence Consider (CIII) response was inversely related to the number of NNRTI Virological failure Consider (CIII) Malabsorption Recommended (BIII) mutations, with no single mutation predicting high-level Unusual or unlicensed dosing Consider (CIII). Examples are once-daily resistance. However, a range of double mutations in the regimens boosted (LPV, SQV, APV, IDV) or NNRTI pocket was associated with reduced response, unboosted (NFV) regimens Children Recommended (BIII) particularly when one of these was at codon 181. The side effects were few in these studies apart from an evanescent APV, amprenavir; ATZ, atazanavir; EFV, efavirenz; IDV, indinavir; LPV, rash. At present this drug is given as two pills twice a day lopinavir; NFV, nelfinavir; SQV, saquinavir. Please refer to the full guidelines for further information (HIV Med 2005; 6 although it does have a long half-life and could be given as (Suppl. 2): 1061). four pills once a day. TMC125 will be available on a compassionate release basis in the UK from October 2006. Studies have shown a slow reversal of lipoatrophy when Darunavir (TMC114) d4T and possibly ZDV are substituted by other drugs such as ABC and TDF. There are convincing data to suggest that This PI, which has a similar structure to amprenavir, was avoiding PIs as first line, or switching from them, leads to a developed because of its in vitro activity against viral better lipid profile, a reduction in calculated risk of future mutants resistant to many of the current members of the PI cardiovascular events and possibly a reduction of insulin class. In Phase IIB studies (POWER 1 and 2) which have led resistance. This class effect on lipids and insulin resistance to the licensure of this drug in the USA, TMC114 600 mg does not apply to ATZ, and boosting with ritonavir does twice daily boosted by ritonavir was shown to be the not appear to change this. optimum dose. In these studies of triple-class experienced

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 499

patients, many of whom had viruses resistant to all the tinued at 16 weeks because of a higher virological failure presently available PIs, 46% of patients in the TMC114 arm rate. Studies in treatment-experienced patients with CCR5 had virological undetectability at 48 weeks (o50 copies/ tropic virus are ongoing but a study in those with dual/ mL) compared to 10% in the comparator boosted-PI arm mixed tropic virus (R5 and X4) demonstrated no benefits [88]. The rates of undetectability were related to the over the optimized background arm, although the drug number of sensitive drugs (including T20) that the patients appeared safe and well tolerated. This is reassuring, as it received concomitantly and the phenotypic sensitivity to does not suggest that emergence of potentially more virulent TMC114 at the start of the study. Lipid abnormalities and X4 tropic strains will be a significant problem with this class other side effects were similar to those seen in the of drugs. Dose adjustment with RTV is required. optimized background regimen. On virological failure of The above drugs are likely to find a role in triple-class TMC114, the majority of isolates sensitive to TPV at experienced patients, where the available options are baseline remained sensitive to this drug. Studies in naı¨ve diminishing. In addition, if appropriate studies are carried patients are examining a lower dose of TMC114 given out, these drugs may be used as a substitute for one of the once a day compared with a boosted LPV regimen. TMC114 backbone drugs, particularly after is available on a compassionate release basis in the UK. initial treatment failure when finding a second appropriate nucleoside analogue is often difficult. Integrase inhibitors This novel class of drug interferes with the ability of HIV References to integrate into the human genome, and three different 1 Levy RW, Rayner CR, Fairley CK et al. Multidisciplinary HIV companies (Gilead Sciences, GlaxoSmithKline and Merck) adherence intervention: a randomized study. AIDS Patient Care have drugs in development. The most advanced of these STDs 2004; 18: 728–735. drugs is the Merck product and there are preliminary 2 Rathbun RC, Farmer KC, Stephens JR, Lockhart SM. Impact of results from Phase IIb/III dose-ranging studies in both an adherence clinic on behavioral outcomes and virologic naı¨ve and triple-class experienced patients. In the experi- response in treatment of HIV infection: a prospective, enced patient group, approximately 80% of patients had a randomized, controlled pilot study. Clin Ther 2005; 27: viral load of less than 400 copies/mL by 16 weeks 199–209. compared with 20% in the optimized background-only 3 Remien RH, Stirratt MJ, Dolezal C et al. Couple-focused support group. In the naı¨ve study, high rates of virological to improve HIV medication adherence: a randomized undetectability equal to those observed with EFV were controlled trial. AIDS 2005; 19: 807–814. seen with a TDF/3TC backbone (485% in all arms by 24 4 Gallant JE, DeJesus E, Arribas JR et al. for the Study 934 weeks), with a more rapid decline in viral load in the Group. Tenofovir DF, emtricitabine, and efavirenz vs. integrase arms at weeks 4 and 8. Both studies are ongoing, zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med but so far few side effects have been observed. Durability 2006; 354: 251–260. and long-term toxicity have yet to be defined, but 5 Schackman BR, Finkelstein R, Neukermans CP, Lewis L, Eldred importantly the drug is metabolized by glucuronidination L. The cost of HIV medication adherence support interventions: and does not require RTV boosting. It is hoped that this results of a cross-site evaluation. AIDS Care 2005; 17: drug will become an important new treatment option, and 927–937. it will be available on a compassionate release basis in the 6 MRC Clinical Trial Units. http://www.ctu.mrc.ac.uk/studies/ UK from October 2006. spartac.asp 7 Philips A for the CASCADE Collaboration. Short-term risk of CCR5 inhibitors AIDS according to current CD4 cell count and viral load in One drug in this class (aplaviroc) has been discontinued for antiretroviral drug-naive individuals and those treated in the hepatotoxicity and, with another, suboptimal performance monotherapy era. AIDS 2004; 18: 51–58. in a naı¨ve population was observed because of inadequate 8 Bayer R, Fairchild AL. Changing the paradigm for HIV testing – dosing (). The most advanced drug in this class is the end of exceptionalism. N Engl J Med 2006 355: 647–649. (UK-427857), which is in Phase IIb/III trials in 9 Opravil M, Ledergerber B, Furrer H et al. Clinical efficacy of both experienced and naı¨ve patients. The naı¨ve study early initiation of HAART in patients with asymptomatic compares once-daily (lower dose) UK-427857, twice-daily HIV infection and CD4 cell count 4350106. AIDS 2002 16: UK-427857, and EFV, with a backbone of Combivir in 1371–1381. patients with CCR5 tropic virus. The study is ongoing, 10 Pallela JF, Delona-Knoll M, Cmiel JS et al. Survival benefit of although the lower once-a-day dosing arm was discon- initiating antiretroviral therapy in HIV-infected persons in

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 500 B Gazzard

different CD4 1 cell strata. Ann Intern Med 2003; 138: 320– 22 Gallant JE, Staszewski S, Pozniak A et al. Efficacy and safety of 326. tenofovir DF vs. stavudine in combination therapy in 11 Sterne J, May M, Costagliola D et al. Estimating the antiretroviral-naive patients. A 3-year randomized trial. JAm optimum CD4 threshold for starting HAART in ART-naı¨ve Med Assoc 2004; 292: 191–201. HIV-infected individuals. 13th Conference on Retroviruses and 23 van Leth F, Phanuphak P, Ruxrungtham K et al. Comparison of Opportunistic , Denver, CO, February 2006 first-line antiretroviral therapy with regimens including [Abstract 525]. nevirapine, efavirenz, or both drugs, plus stavudine and 12 Phillips AN, Dunn DT, Sabin C et al. for the UK collaborative lamivudine: a randomised open-label trial, the 2NN Study. Group on HIV Drug Resistance and the UK CHIC Study Group. Lancet 2004; 363: 1253–1263. Long term probability of detection of HIV-1 drug resistance 24 van Leth F, Andrews S, Grinsztejn B et al. The effect of baseline after starting antiretroviral therapy in routine clinical practice. CD4 cell count and HIV-1 viral load on the efficacy and safety AIDS 2005; 19: 487–494. of nevirapine or efavirenz-based first-line HAART. AIDS 2005; 13 Harrigan PR, Hogg RS, Dong WW et al. Predictions of HIV 19: 463–471. drug-resistance mutations in a large antiretroviral-naive 25 Boly L, Cafaro V, Dyner T. Depressive symptoms predict cohort initiating triple antiretroviral therapy. J Infect Dis 2005; increased incidence of neuropsychiatric side-effects in patients 191: 339–347. treated with efavirenz. J Acquir Immune Defic Syndr 2006; 42: 14 El-Sadr W, Neaton J. Episodic CD4-guided use of ART is 514–515. inferior to continuous therapy: results of the SMART Study. 26 Important change in SUSTIVA (efavirenz) package insert: 13th Conference on Retroviruses and Opportunistic Infections, change from pregnancy category C to D. Bristol-Myers Squibb Denver, CO, February 2006 [Abstract 106LB]. Company, March 2005. 15 Gallant JE, DeJesus E, Arribas JR et al. for the Study 934 27 Montaner JS, Cahn P, Zala C et al. Randomized, controlled Group. Tenofovir DF emtricitabine, and efavirenz vs. study of the effects of a short course of prednisone on the zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med incidence of rash associated with nevirapine in patients 2006; 354: 251–260. infected with HIV-1. J Acquir Immune Defic Syndr 2003; 33: 16 Montaner JS, Sang MS, Barylski C et al. FOCUS Study: 41–46. saquinavir qd regimen versus efavirenz qd regimen 48 week 28 Squires K, Lazzarin A, Gatell JM. Comparison of once-daily analysis in HIV infected patients. 42nd Interscience Conference atazanavir with efavirenz, each in combination with fixed-dose on Antimicrobial Agents and Chemotherapy. San Diego, CA, zidovudine and lamivudine, as initial therapy for patients September 2002 [Abstract H-167]. infected with HIV. J Acquir Immune Defic Syndr 2004; 36: 17 Shafer RW, Smeaton LM, Robbins GK et al. Comparison of 1011–1019. four-drug regimens and pairs of sequential three-drug 29 Staszewski S, Keiser P, Montaner J et al. Abacavir- regimens as initial therapy for HIV-1 infection. N Engl J Med lamivudine-zidovudine vs indinavir-lamivudine-zidovudine 2003; 349: 2304–2315. in antiretroviral-naive HIV-infected adults: a randomized 18 Bartlett JA, Johnson J, Herrera G et al. Long-term results of equivalence trial. J Am Med Assoc 2001; 285: initial therapy with abacavir and lamivudine combined with 1155–1163. efavirenz, amprenavir/ritonavir or stavudine. J Acquir Immune 30 Walmsely S, Bernstein B, King M et al. M98-863 Study Team. Defic Syndr 2006; 43: e-pub ahead. Lopinavir-ritonavir versus nelfinavir for the initial treatment of 19 Riddler SA, Haubrich R, DiRienzo G et al. A prospective, HIV infection. N Engl J Med 2002; 346: 2039–2046. randomized, phase III trial of NRTI-, PI-, and NNRTI-sparing 31 Gathe JC Jr, Ive P, Wood R et al. SOLO: 48-week efficacy and regimens for initial treatment of HIV-1 infection: ACTG 5142. safety comparison of once daily fosamprenavir/ritonavir versus XVI International AIDS Conference. Toronto, Canada, August twice-daily nelfinavir in naı¨ve HIV-1-infected patients. AIDS 2006. [Abstract THLB0204]. 2004; 18: 1529–1537. 20 Eron J, Yeni P, Gathe J et al. The KLEAN study of 32 Dragsted U, Gerstoft J, Pedersen C et al. Randomized trial to fosamprenavir-ritonavir versus lopinavir-ritonavir, each in evaluate indinavir/ritonavir versus saquinavir/ritonavir in combination with abacavir-lamivudine, for initial treatment of human immunodeficiency type-1 infected patients: the HIV infection over 48 weeks: a randomised non-inferiority MaxCMin1 trial. J Infect Dis 2003; 188: 634–642. trial. Lancet 2006; 368: 476–482. 33 Dragsted UB, Gerstoft J, Youle M et al. A randomized trial to 21 Staszewski S, Morales JM, Tashima KT et al. Efavirenz evaluate lopinavir/ritonavir versus saquinavir/ritonavir in plus zidovudine and lamivudine, efavirenz plus indinavir, HIV-1-infected patients: the MaxCmin2 trial. Antivir Ther and indinavir plus zidovudine and lamivudine in the treatment 2005; 10: 735–743. of HIV-1 infection in adults. N Engl J Med 1999; 341: 34 Murphy RL, Sanne I, Cahn P et al. Dose ranging, randomized, 1865–1873. clinical trial of atazanavir with lamivudine and stavudine in

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 501

antiretroviral naı¨ve subjects: 48 week results. AIDS 2003; 17: 45 Jones R, Stebbing J, Nelson M et al. Renal dysfunction with 2603–2614. fumarate-containing highly active 35 Sanne I, Piliero P, Squires K et al. Results of a phase 2 clinical antiretroviral is not observed more frequently: a cohort and trial at 48 weeks (AI424-067): a dose-ranging, safety, and case-controlled study. J Acquir Immune Defic Syndr 2004; 37: efficacy comparative trial of atazanavir at three doses in 1489–1495. combination with didanosine and stavudine in anti- 46 Nelson M, Cooper D, Schooley R et al. The safety of tenofovir retroviral naı¨ve subjects. J Acquir Immune Defic Syndr 2003; DF for the treatment of HIV infection: the first 4 years. 13th 32: 18–29. Conference on Retroviruses and Opportunistic Infections. 36 Squires K, Lazzarin A, Gatell JM et al. Comparison of once- Denver, CO, February 2006 [Abstract 781]. daily atazanavir with efavirenz, each in combination with 47 Moreno S, Domingo P, Palacios R et al. Renal safety of fixed-dose zidovudine and lamivudine, as initial therapy for tenofovir disoproxil fumarate in HIV-1 treatment-experienced patients infected with HIV. J Acquir Immune Defic Syndr 2004; patients with adverse events related to prior NRTI use: data 36: 1011–1019. from a prospective, observational, multicentre study. J Acquir 37 Malan N, Krantz E, David N et al. Efficacy and safety of Immune Defic Syndr 2006; 42: 384–385. atazanavir-based therapy in antiretroviral naive HIV-1 infected 48 Gallant JE, Parish MA, Keruly JC et al. Changes in renal subjects, both with and without ritonavir: 48-week results from function associated with tenofovir disoproxil fumarate, AI424-089. 13th Conference on Retroviruses and Opportunistic compared with nucleoside reverse transcriptase inhibitor Infections. Denver, CO, February 2006 [Abstract 107LB]. treatment. Clin Infect Dis 2005; 40: 1194–1198. 38 Gallant JR, Rodriguez AE, Weinberg WG et al. ESS30009 study. 49 Winston A, Amin J, Mallon P et al. Minor changes in calculated Early virologic nonresponse to tenofovir, abacavir and creatinine clearance and anion gap are associated with lamivudine in HIV-infected antiretroviral-naı¨ve subjects. tenofovir disoproxil fumarate-containing highly active J Infect Dis 2005; 192: 1921–1230. antiretroviral therapy. HIV Med 2006; 7: 105–111. 39 Jemsek J, Hutcherson P, Harper E. Poor virological responses 50 Johnson MA, Gathe JC Jr, Podzamczer D et al. A once-daily and early emergence of resistance in treatment naı¨ve, HIV- lopinavir/ritonavir-based regimen provides noninferior infected patients receiving a once daily triple nucleoside antiviral activity compared with a twice-daily regimen. regimen of didanosine, lamivudine, and tenofovir DF. 11th J Acquir Immune Defic Syndr 2006; 43: 153–160. Conference on Retroviruses and Opportunistic Infections, San 51 Rauch A, Nolan D, Martin A, McKinnon E, Almeida C, Mallal S. Francisco, CA, 2004 [Abstract 51]. Prospective genetic screening decreases the incidence of 40 Curtis H, Sabin CA, Johnson MA. Findings from the first abacavir hypersensitivity reaction in the Western Australian national clinical audit of treatment for people with HIV. HIV HIV cohort study. Clin Infect Dis 2006; 43: 99–102. Med 2003; 4: 11–17. 52 Reeves I, Churchill D, Fisher M. Clinical Utility of HLA-B*5701 41 DeJesus E, Herrera G, Teofilo E et al. Abacavir versus testing in a UK clinic cohort. 12th Annual BHIVA Conference, zidovudine combined with lamivudine and efavirenz, for the Brighton, UK, March 2006 [Abstract 019]. treatment of antiretroviral-naı¨ve HIV-infected adults. Clin 53 Podzamcer D, Ferrer E, Sa´nchez P et al. A randomized Infect Dis 2004; 39: 1038–1046. comparison of abacavir and stavudine, combined with 42 Gallant JE, Pozniak AL, DeJesus E et al. Efficacy and safety of lamivudine/efavirenz, in antiretroviral-naive patients. Final tenofovir DF (TDF), emtricitabine (FTC) and efavirenz (EFV) 96-week results (ABCDE study). 12th Conference on compared to fixed dose zidovudine/lamivudine (CBV) and EFV Retroviruses and Opportunistic Infections. Boston, MA, through 96 weeks in antiretroviral treatment-naive patients. February 2005 [Abstract 587]. XVI International AIDS Conference. Toronto, Canada, August 54 Maitland D, Moyle G, Hand J et al. Early virologic failure in 2006. [Abstract TUPE0064]. HIV-1 infected subjects on didanosine/tenofovir/efavirenz: 43 Moyle G, DeJesus E, Cahn P et al. Abacavir once or twice 12-week results from a randomized trial. AIDS 2005; 19: daily combined with once-daily lamivudine and efavirenz 1183–1188. for the treatment of antiretroviral-naive HIV-infected 55 Saag MS, Cahn P, Raffi F et al. Efficacy and safety of adults: results of the ziagen once daily in antiretroviral emtricitabine vs. stavudine in combination therapy in combination Study. J Acquir Immune Defic Syndr 2005; 38: antiretroviral-naive patients. A randomized trial. JAmMed 417–425. Assoc 2004; 292: 180–190. 44 Sosa N, Hill-Zabala C, DeJesus E et al. Abacavir and 56 Toro, Hicks C and the RESIST-1 Team. RESIST-1: A phase 3 lamivudine fixed dose combination once daily compared with randomized, controlled, open-label multicenter trial comparing abacavir and lamivudine twice daily in HIV infected patients tipranavir/ritonavir (TPV/r) to an optimized comparator over 48 weeks. J Acquir Immune Defic Syndr 2005; 40: protease inhibitor/r (CPI/r) regimen in antiretroviral (ARV) 422–427. experienced patients: 24-week data. 44th Interscience

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 502 B Gazzard

Conference on Antimicrobial Agents and Chemotherapy. 68 Machouf N, Thomas R, Nguyen VK et al. Effects of drug Washington DC, October–November 2004 [Abstract H-1137a]. resistance on viral load in patients failing antiretroviral 57 Cahn P and the RESIST 2 Team. 24 week data from RESIST 2: therapy. J Med Virol 2006; 78: 608–613. phase 3 study of the efficacy and safety of either tipranavir/ 69 Nijhuis M, Schuurman R, de Jong D et al. Increased fitness of ritonavir or an optimized ritonavir-boosted standard of care drug resistant HIV-1 protease as a result of acquisition of comparator PI in a large randomized multicenter trial in compensatory mutations during suboptimal therapy. AIDS treatment-experienced HIV 1 patients. 7th International 1999; 13: 2349–2359. Congress on Drug Therapy in HIV Infection. Glasgow, UK, 70 Palmer S, Kearney M, Maldarelli F. Multiple, linked human November 2004 [Abstract PL14.3]. immunodeficiency virus type 1 drug resistance mutations 58 Katlama C, Carvalho MI, Cooper D et al. TMC 114/r out in treatment-experienced patients are missed by performs investigator selected PI(s) in three class-experienced standard genotype analysis. J Clin Microbiol 2005; 43: patients: week 24 primary analysis of POWER 1 (TMC 114- 406–413. C213). 3rd IAS Conference on HIV Pathogenesis and Treatment. 71 Deeks SG, Wrin T, Liegler T et al. Virologic and immunologic Rio de Janeiro, Brazil, July 2005 [Abstract WeOaLB0102]. consequences of discontinuing combination antiretroviral- 59 Thommes JA, Demasi R, Haubrich R. Improved virologic drug therapy in HIV patients with detectable viremia. N Engl response in three-class experienced patients when an active J Med 2001; 344: 472–480. boosted protease inhibitor is used with enfuvirtide (ENF). 43rd 72 Lawrence J, Huppler Hullsiek K, Thackeray L et al. Final results Annual Infectious Disease Society of America. San Francisco, of CPCRA 064: a randomized trial examining structured CA, September 2005 [Abstract 785]. treatment interruption for patients failing therapy with multi- 60 Hill A, Moyle G. Relative antiviral efficacy of TMC114/r and drug resistant HIV. 12th Conference on Retroviruses and tipranavir/r versus control PI in the POWER and RESIST trials. Opportunistic Infections. Boston, MA, February 2005 [Abstract 12th Annual Conference of the British HIV Association. 579]. Brighton, UK, March 2006 [Abstract P1]. 73 Walmsley S, LaPierre N, Loutfy M et al. CTN 164: a prospective 61 Murray JS, Elashoff MR, Iacono-Connors LC et al. The use of randomized trial of structured treatment interruption vs plasma HIV RNA as a study endpoint in efficacy trials of immediate switching in HIV-infected patients experiencing antiretroviral drugs. AIDS 1999; 13: 797–804. virologic failure on HAART. 12th Conference on Retroviruses 62 Miller V, Sabin C, Hertogs K et al. Virological and and Opportunistic Infections. Boston, MA, February 2005 immunological effects of treatment interruptions in HIV-1 [Abstract 580]. infected patients with treatment failure. AIDS 2000; 14: 74 Beatty G, Lu J, Hunt P et al. Randomized pilot study of 2857–2867. immediate enfuvirtide-based therapy vs a treatment 63 Ledergerber V, Lundgren JD, Walker AS et al. of trend in CD4- interruption followed by enfuvirtide-based therapy in highly positive T-cell count and mortality among HIV-1-infected treatment-experienced patients. Abstracts of the 12th individuals with virological failure to all three antiretroviral- Conference on Retroviruses and Opportunistic Infections. drug classes. Lancet 2004; 364: 51–62. Boston, MA, February 2005 [Abstract 581]. 64 Raffanti SP, Fusco JS, Sherrill BH et al. Effect of persistent 75 Deeks SG, Hoh R, Neilands TB et al. Interruption of treatment moderate viremia on disease progression during with individual therapeutic drug classes in adults with HIV therapy. J Acquir Immune Defic Syndr 2004; 37: multidrug-resistant HIV-1 infection. J Infect Dis 2005; 192: 1147–1154. 1537–1544. 65 Solomon A, Lane N, Wightman F, Gorry PR, Lewin SR. 76 Dragsted U, Fox Z, Mathiesen L et al. for the COLATE Trial Enhanced replicative capacity and pathogenicity of HIV-1 Group. Final week 48 analysis of a phase 4, randomised, open- isolated from individuals infected with drug-resistant virus and label, multi-center trial to evaluate safety and efficacy of declining CD4 1 T-cell counts. J Acquir Immune Defic Syndr continued lamivudine twice daily versus discontinuation of 2005; 40: 140–148. lamivudine in HIV-1-infected adults with virological failure on 66 Kristiansen TB, Pedersen AG, Eugen-Olsen J, Katzenstein TL, ongoing combination treatments containing lamivudine: The Lundgren JD. Genetic evolution of HIV in patients remaining COLATE Trial. 11th Conference on Retroviruses and on a stable HAART regimen despite insufficient viral Opportunistic Infections. San Francisco, CA, February 2004 suppression. Scand J Infect Dis 2005; 37: 890–901. [Abstract 549]. 67 Hatano H, Hunt P, Weidler J et al. Rate of viral evolution and 77 Eron JJ Jr, Bartlett JA, Santana JL et al. Persistent antiretroviral risk of losing future drug options in heavily pre-treated activity of nucleoside analogues after prolonged zidovudine patients remaining on stable partially suppressive regimen. and lamivudine therapy as demonstrated by rapid loss of 13th Conference on Retroviruses and Opportunistic Infections. activity after discontinuation. J Acquir Immune Defic Syndr Denver, CO, February 2006 [Abstract 615]. 2004; 37: 1581–1583.

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503 BHIVA guidelines 503

78 Campbell TB, Shulman NS, Johnson SC et al. Antiviral activity a French nationwide study. 4th European HIV Drug of lamivudine in salvage therapy for multidrug-resistant HIV-1 Resistance Workshop. Monte Carlo, Monaco, March 2006 infection. Clin Infect Dis 2005; 41: 236–242. [Abstract 6]. 79 Castagna A, Danise A, Menzo S et al. Lamivudine monotherapy 84 Castagna A, Danise A, Menzo S et al. Lamivudine monotherapy in HIV-1-infected patients harbouring a lamivudine-resistant in HIV-1-infected patients harbouring a lamivudine-resistant virus: a randomized pilot study (E-184 V study). AIDS 2006; virus: a randomized pilot study (E-184 V study). AIDS 2006; 20: 795–803. 20: 795–803. 80 Wensing AMJ et al. First representative prospective 85 Stone C, Holbrook J, Madsen H, Modha S, Craig C. surveillance data on HIV baseline drug resistance from 17 Amplification and sequencing of HIV-1 from low viral load countries in Europe; the SPREAD-programme. 4th European plasma samples and analysis and interpretation of sequences HIV Drug Resistance Workshop. Monte Carlo, Monaco, March obtained. 4th European HIV Drug Resistance Workshop. Monte 2006 [Abstract 1]. Carlo, Monaco, March 2006 [Abstract 63]. 81 Cane P, Chrystie I, Dunn D et al. Time trends in primary 86 Johnson JA, Li J-F, Wei X et al. Baseline detection of low- resistance to HIV drugs in the United Kingdom: multicentre frequency drug resistance-associated mutations is strongly observational study. Br Med J 2005; 331: 1368–1372. associated with virological failure in previously antiretroviral- 82 Garcia-Diaz A, Booth C, Nebbia G, Chawla A, Johnson M, naı¨ve HIV-1-infected persons. XV International HIV Drug Geretti AM. Transmitted drug resistance clusters with the Resistance Workshop. Sitges, Spain, July 2006 [Abstract 46]. infecting HIV-1 subtype: a single centre analysis of all new 87 Geretti AM. HIV-1 subtypes: epidemiology and significance for HIV-1 diagnoses in London. XV International Drug Resistance HIV management. Curr Opin Infect Dis 2006; 19: 1–7. Workshop. Sitges, Spain, June 2006 [Abstract 113]. 88 Lazzarin A, Queiroz-Telles F, Frank I et al. TMC114 provides 83 Costagliola D, Descamps D, Assoumou L et al. and the durable viral load suppression in treatment-experienced ANRS AC11 Resistance Study Group. Prevalence of patients: POWER 1 and 2 combined week 48 analysis. resistance to at least one drug in treated HIV infected XVI International AIDS Conference, Toronto, Canada, August patients with viral load 41,000 copies/mL in 2004: 2006 [Abstract TUAB0104].

r 2006 British HIV Association HIV Medicine (2006) 7, 487–503